Digital White Papers

I L T A W H I T E P A P E R | L I T I G A T I O N A N D P R A C T I C E S U P P O R T 38 Horizontal measurement allows you to determine whether there are missing and/ or misrepresentative fields. You may have an expectation that certain information is available in your data or that a field represents information based on its field name. For this measurement, you review the data for completeness by assessing the available fields against a list of expected fields. Consider a mobile device forensic collection: You may expect to find chat application data alongside call, SMS, address book, and other information— each with a specific set of fields. The horizontal measurement would consist of evaluating each set of fields and determine whether the relevant sets of information data (e.g., chat application data) and assigned fields (e.g., sender username) are all present. In the vertical measurement, you are determining the level of completeness of the records. The primary tests for these are to compare totals in the data against a control total, such as record counts or another aggregated sum. This measurement can also be performed by testing distributions. Examining the total records in a time series helps identify date period gaps. The measurement can also be performed by other groupings, such as financial account numbers or custodians. The inter-data measurement compares different data sets to determine whether there are any differences within your data as a whole. This can take many different forms, for it is dependent on the structure of your data and your requirements. Even though a data set may appear to be complete on its own, comparing it to another data set on a related field or set of fields may expose data gaps. This test allows you to see how complete your entire corpus of data is in relation to each data set. Measuring Validity and Reliability Measuring data validity and reliability helps you gauge how meaningful data is. The goal of any analysis is to accurately represent information. If your analysis is built on misrepresentative data, or the analysis does not accurately present the data, the analysis is useless—or worse, it is misleading and will yield bad results. The challenge is designing metrics to effectively measure meaning within data. In addition to determining how complete your data is, you can also measure whether you have true and valid data. Not all data sets will reflect information the way you might want. The systems and processes that generate data changes over time. Misconfigurations of a system, changes to data input forms, differences in how individuals created the data, and other factors can all affect the quality of your data. On top of that, data from two different sources might represent similar information

• Cover